Cactaceae) Paraphyly for the Transition to the Cactus Life Form1

Home , Cactoideae, Maihuenia poeppigii, Pereskiopsis

American Journal of Botany 92(7): 1177±1188. 2005.

BASAL CACTUS PHYLOGENY: IMPLICATIONS OF PERESKIA (CACTACEAE) PARAPHYLY FOR THE TRANSITION TO THE CACTUS LIFE FORM1

ERIKA J. EDWARDS,2,4 RETO NYFFELER,3 AND MICHAEL J. DONOGHUE2

2Department of Ecology and Evolutionary Biology and the Peabody Museum of Natural History, Yale University, P.O. Box 208105, New Haven, Connecticut 06520-8105 USA; and 3Institut fuÈr Systematische Botanik, University of Zurich, Zollikerstrasse 107, CH-8008 Zurich, Switzerland

The cacti are well-known desert plants, widely recognized by their specialized growth form and essentially lea¯ess condition. Pereskia, a group of 17 species with regular leaf development and function, is generally viewed as representing the ``ancestral cactus,'' although its placement within Cactaceae has remained uncertain. Here we present a new hypothesis of phylogenetic relationships at the base of the Cactaceae, inferred from DNA sequence data from ®ve gene regions representing all three plant genomes. Our data support a basal split in Cactaceae between a clade of eight Pereskia species, centered around the Caribbean basin, and all other cacti. Two other Pereskia clades, distributed mostly in the southern half of South America, are part of a major clade comprising Maihuenia plus Cactoideae, and Opuntioideae. This result highlights several events in the early evolution of the cacti. First, during the transition to stem-based photosynthesis, the evolution of stem stomata and delayed bark formation preceded the evolution of the stem cortex into a specialized photosynthetic tissue system. Second, the basal split in cacti separates a northern from an initially southern cactus clade, and the major cactus lineages probably originated in southern or west-central South America.

Key words: biogeography; Cactaceae; inferior ovary; Pereskia; phylogeny; phytochrome C; stem-based photosynthesis; trnK/ matK.

The Cactaceae contain 1500±1800 species renowned for genetic studies support the monophyly of the opuntioid and their remarkable morphological and physiological adaptations cactoid subfamilies, but their relationships to Maihuenia and to drought (Barthlott and Hunt, 1993). Although they are Pereskia have remained unresolved. This is due both to limited found in a range of environments, they are especially con- sampling of Pereskia (Hershkovitz and Zimmer, 1997; Apple- spicuous components of the New World's arid regions and quist and Wallace, 2001; Nyffeler, 2002), as well as an in- represent one of the world's most spectacular desert radiations. ability of the molecular data to resolve basal cactus nodes The great majority of cactus diversity is found in two major (Wallace, 1995). To date, no studies have con®rmed or rejected lineages, the Opuntioideae and Cactoideae. Most members of the monophyly of Pereskia. these groups are what might be regarded as ``typical'' cacti: Pereskia species are often interpreted as ``relictual cacti,'' they are stem succulents with only vestigial or ephemeral and are used as a general model of the ancestral condition from leaves and a well-developed photosynthetic stem cortex with which the highly specialized morphology and physiology of CAM carbon metabolism, specialized ``collapsable'' xylem the core cacti arose (Rauh, 1979; Gibson and Nobel, 1986; cells that aid in water storage (Schleiden, 1845; Mauseth et Mauseth and Landrum, 1997). Pereskia species are widely dis- al., 1995), deeply recessed inferior ovaries, and specialized tributed in the Caribbean and Central and South America in a short shoots (areoles) with very reduced internodes that pro- range of drier forest habitats. They have been described as duce spines, new long shoots, glochids (in Opuntioideae), and having superior to inferior ovaries, broad, ¯attened leaves with ¯owers. The remaining cacti consist of two small genera, Per- C3 photosynthesis, areoles with leaf production, dense, ®brous eskia and Maihuenia. These two were at one time united into wood, a simple cortex without cortical bundles, poorly devel- a third subfamily, Pereskioideae, but this grouping is supported oped stem epidermal and hypodermal layers, nonsucculent tis- only by the shared absence of many of the typical cactus characters noted above, and in recent years Maihuenia has been sues, and as inhabiting relatively mesic environments (Mau- placed in its own subfamily, Maihueniodeae. Multiple phylo- seth and Landrum, 1997). This generalized depiction of Per- eskia species has led botanists to believe that the stem suc- 1 Manuscript received 7 October 2004; revision accepted 8 April 2005. culent cacti are derived from woody, nonsucculent trees with The authors thank Wendy Applequist, Miriam Diaz, Urs Eggli, Beat Leuen- C3 photosynthesis, as opposed to other growth forms (e.g., an berger, the Desert Botanical Garden (Phoenix, Arizona USA), the Berlin Bo- herbaceous, succulent CAM plant) (Grif®th, 2004). tanical Garden (Berlin-Dahlem, Germany), the Sukkulenten-Sammlung (ZuÈ- While the ``Pereskia model'' has been useful, it also down- rich, Switzerland), and the JardõÂn BotaÂnico Nacional (Santo Domingo, Do- plays some potentially important ways in which Pereskia spe- minican Republic) for tissue or DNA samples and/or logistical support in the ®eld. We owe special thanks to Dianella Howarth and Philippe Reuge for cies differ from one another. An alternative perspective of Per- their help and advice in the lab and to Casey Dunn, Nico Cellinese, and the eskia is supported by other studies that have emphasized the Donoghue Lab for helpful discussions. Beat Leuenberger and two anonymous substantial ecological, morphological, and anatomical diversity reviewers provided comments that greatly improved this manuscript. This found within Pereskia (Schumann, 1898; Berger, 1926, 1929; work was funded in part from a National Science Foundation Graduate Re- Boke, 1954; Backeberg, 1958; Bailey, 1962; Bailey, 1963a, b, search Fellowship and a Deland Award for Student Research from the Arnold Arboretum of Harvard University to E. J. E, and by a grant from the Swiss c, d; Boke, 1963, 1966, 1968; Leuenberger, 1986; Martin and National Science Foundation (823A-056624) to R. N. Wallace, 2000). In his monograph of the genus, Leuenberger 4 Author for correspondence (e-mail: [email protected]) (1986) delineated several species groups within Pereskia that 1177 1178 AMERICAN JOURNAL OF BOTANY [Vol. 92

TABLE 1. Pereskia species groups and distribution of traits, as outlined by Leuenberger (1986).

Delayed bark No areole leaf Species formation Stem stomata production Tuberous roots Inferior ovary Pereskia aculeata ÐX X Ð Ð Pereskia grandifolia XXÐ Ð Ð Pereskia bahiensis XXÐ Ð Ð Pereskia sacharosa XXÐ Ð Ð Pereskia stenantha XXÐ Ð Ð Pereskia nemorosa XÐÐ Ð Ð Pereskia horrida XXX X Ð Pereskia diaz-romeroana XXX X Ð Pereskia weberiana XXX X Ð Pereskia bleo ÐÐ X Ð X Pereskia quisqueyana ÐÐ X X X Pereskia portulacifolia ÐÐÐ X X Pereskia marcanoi ÐÐÐ X X Pereskia zinnii¯ora ÐÐÐ X X Pereskia lychnidi¯ora ÐÐÐ Ð Ð Pereskia guamacho ÐÐÐ X Ð Pereskia aurei¯ora ÐÐÐ Ð Ð

are marked by particular combinations of vegetative and re- for 2 min, ®nal extension of 72ЊC for 10 min; psbA-trnH:95ЊC for 2 min, 35 productive features (Table 1), but gave no indication of how cycles of 94ЊC for 30 s, 50ЊC for 30 s, 72ЊC for 1 min, ®nal extension of these subgroups might be related to one another, nor did he 72ЊC for 7 min; cox3: 98ЊC for 5 min, 40 cycles of 95ЊC for 1 min, 56ЊC for explicitly argue that Pereskia is monophyletic. Our ability to 1 min, 72ЊC for 1 min 30 s, ®nal extension of 72ЊC for 5 min; phyC:94ЊC infer early events in the evolutionary history of the cacti rests for 5 min, then a ``touchdown'' protocol with a progressive drop in annealing squarely on resolving two outstanding problems in cactus phy- temperature every 3 cycles (94ЊC for 1 min, 68ЊC±65ЊC±62ЊC±59ЊC±56ЊC for logeny: (1) whether Pereskia is monophyletic, and (2) how 2 min, 72ЊC for 2 min), 25 cycles of 94ЊC for 1 min, 53ЊC for 1 min, 72ЊC Њ Њ Pereskia species are related to the rest of the cacti. for 1 min, ®nal extension of 72 C for 15 min; trnK/matK:95C for 4 min, 40 cycles of 94ЊC for 30 s, 48±52ЊC for 1 min, 72ЊC for 1 min 30 s, ®nal To address these questions, we obtained new DNA sequenc- Њ es from all species of Pereskia and Maihuenia and from rep- extension of 72 C for 7 min. PCR products were cleaned using a PCR Puri- resentatives of Opuntioideae, Cactoideae, and selected Portu- ®cation kit (Qiagen) and directly sequenced, with the exception of the phytochrome C region. PhyC PCR product was cloned using the Invitrogen TOPO lacaceae. We selected ®ve gene regions representing the three TA Cloning Kit for Sequencing (Carlsbad, California USA), and 2±6 clones plant genomes: the nuclear gene phytochrome C, rbcL, trnK/ of appropriate size (ϳ1.1 kb) per taxon were sequenced. Except as noted matK, and psbA-trnH IGS from the chloroplast genome, and later, clones from each taxon were nearly identical and formed monophyletic the mitochondrial gene cox3. Based on phylogenetic analyses groups in a parsimony analysis of all clones (tree not shown), so we conclude of these data, we present a new hypothesis of basal relation- that there is one copy of phyC and that slight differences between clones were ships in Cactaceae and highlight how these ®ndings bear on due to PCR error or multiple alleles. One clone for each species was randomly our understanding of early cactus evolution. We focus specif- chosen for inclusion in the phylogenetic analysis. Two distinct copies of phyC ically on historical biogeography and the earliest steps in the were recovered from Pereskiopsis and Quiabentia, but in each case the ad- transition to stem-based photosynthesis. ditional copy was easily distinguished by its great divergence from the other phyC sequences. MATERIALS AND METHODS Dye terminator cycle sequencing reactions for all regions were performed in 20 ␮L volumes (1.5 ␮L5ϫ buffer, 1±3 ␮L PCR product, 2 ␮L BigDye ␮ Sequence generationÐTotal genomic DNA was isolated from fresh or v.3, 0.5 pmol primer, 13.3±15.3 L ddH2O) using the following protocol: dried leaf or cortical tissue using a slightly modi®ed procedure of the DNeasy 96ЊC for 30 s, 26 cycles of 96ЊC for 10 s, 50ЊCfor5s,60ЊC for 4 min. Plant Mini kit (Qiagen, Valencia, California USA) as described in Nyffeler Reactions were then cleaned using EdgeBiosystem plates (Gaithersburg, (2002). In some instances, extractions were so mucilaginous that the precip- Maryland, USA) and run on either an MJResearch BaseStation51 (MJ Re- itation step was repeated. search, South San Francisco, California, USA) or ABIPrism 3700 automated Five distinct gene regions representing all three plant genomes were am- sequencer (Applied Biosystems). pli®ed and sequenced using standard primers found in the literature (rbcL: Contiguous sequences were constructed and edited using Sequencher v.4.2 Bousquet et al., 1992; trnK/matK: Johnson and Soltis, 1994; Nyffeler, 2002; (Gene Codes Corp., Ann Arbor, Michigan, USA), and homologous sequences psbA-trnH IGS: Sang et al., 1997; phyC: Mathews and Donoghue, 1999; cox3: for each region were aligned in MacClade v. 4.06 (Maddison and Maddison, Duminil et al., 2002) with the exception of phyC, for which two new internal 2003). In most cases, alignment was unambiguous. However, the psbA-trnH sequencing primers were designed (phyC454R: 5Ј GSACCCATGTTTGCC 3Ј IGS region varied considerably in length between taxa, and attempts to align and phyC701F: 5Ј GGATTCAAACTGTGC 3Ј). Fragments were ampli®ed in outgroup sequences with Cactaceae sequences for this region gave dubious 25 ␮L reactions (1 ␮L genomic DNA, 0.2 ␮L AmpliTaq polymerase [Applied results. We therefore coded outgroup taxa as missing for this region. Aligned Biosystems, Foster City, California USA], 2.5 ␮L, 2.5 mmol dNTP, 2.5 ␮L data matrices for all analyses are available from the ®rst author upon request, ϫ ␮ ␮ ␮ 10 buffer, 2.5 L 10 mmol MgCl2, 2.5 L 10 mmol primer, 1±2 L BSA, and are also available from TreeBASE (http://www.treebase.org). ␮ 9.5±11.5 L ddH20) using an automated thermal cycler and standard PCR protocols, with PCR cycling times as follows for the different gene regions: Phylogenetic analysesÐWe arranged the data into two major partitions, rbcL:95ЊC for 5 min, 40 cycles of 94ЊC for 1 min, 48±52ЊC for 1 min, 72ЊC nuclear (phyC) and chloroplast/mitochondrial (rbcL, psbA-trnH IGS, trnK/ July 2005] EDWARDS ET AL.ÐBASAL CACTUS PHYLOGENY 1179

TABLE 2. Sequence information for the different gene partitions.

Sequence phyC psbA/trnH IGS trnK/matK rbcL cox3 Combined Genome Nuclear Chloroplast Chloroplast Chloroplast Mitochondrial Ð Number of taxa in data set 34 29 36 36 36 38 Length of aligned matrix 1133 468 2527 1431 592 6151 Number of constant sites 733 323 2017 1254 563 4890 Number of informative sites 211 63 237 79 13 603 % informative sites 18.6% 13.5% 9.4% 5.5% 2.2% 9.8%

matK, cox3), and analyzed them separately and in combination. We combined its branch lengths. For each simulated data set Di, two ML heuristic searches the cpDNA and mtDNA markers into one partition as both of these organelles were performed, one unconstrained (LML) and one constrained to topology ␦i ϭ i Ϫ i are typically maternally inherited. Also, there is little variation in the mtDNA T0(LT), and the test statistic L ML L T was calculated. The alternative ␦ marker (13 informative sites; see Table 2), and each informative site supports hypothesis T0 was rejected if the original fell outside the 95th percentile of a bipartition that is also supported by the cpDNA markers. We tested for the distribution of ␦i. incongruence between the partitions using a partition homogeneity test (the In performing these analyses, we were concerned that our original data set ILD test of Farris et al., 1994), as implemented in PAUP* version 4.b10 contained missing data, while our simulated data sets did not; the simulated (Swofford, 2002). We then performed parsimony (MP), maximum likelihood data sets might therefore contain more phylogenetic information than our own. (ML), and Bayesian phylogenetic analyses on all three data sets (nuclear, For this reason, we trimmed our combined data set into a smaller matrix with cpDNA/mtDNA, combined) using similar methods for each data set. All MP minimal missing data: in all, we removed six taxa (Ceraria fruiticulosa, Por- and ML analyses were performed using PAUP* version 4.b10; Bayesian anal- tulacaria afra, Pereskiopsis porteri, Quiabentia verticillata, Pereskia diaz- yses used MrBayes v. 3.0b4 (Huelsenbeck and Ronquist, 2001). romeroana, and Pereskia bahiensis), the psbA-trnH IGS region, and the 5Ј trnK intron. This resulted in a data set of length 5060 characters, 450 of which MP analysesÐHeuristic searches were performed using a starting tree built are informative. Modeltest chose the same model (GTR ϩ G ϩ I) with slight by stepwise addition with 1000 random addition replicates and tree-bisection- differences in base frequencies, the substitution rate matrix, and the gamma reconnection (TBR) branch swapping. To assess con®dence in clades, boot- distribution. The ML tree for this data set is topologically identical to the ML strap tests were performed using 1000 bootstrap replicates with 10 random tree for the larger data set. With this smaller data set, we re-ran the parametric addition replicates per bootstrap. bootstrap tests as described.

ML analysesÐModels of molecular evolution were chosen for each data Outgroup jackkni®ngÐTo assess the degree to which outgroup taxon sam- set using Modeltest version 3.05b (Posada and Crandall, 1998). Heuristic pling might affect the position of the root of Cactaceae, we generated a set searches were performed using a starting tree built by stepwise addition with of random combinations of our seven outgroups (both in the number and 100 random addition replicates and TBR branch swapping, and 500 bootstrap identity of taxa) and ran MP analyses using each set. To generate the outgroup replicates with ®ve random addition replicates per bootstrap were performed combinations, a small Perl script was written that is executable in Mac OSX to estimate support values for clades. In the combined data set, only 350 terminal window or Linux (available from the ®rst author or as a free down- bootstrap replicates were performed. load called ``hotgroup'' at http://www.yphy.org/phycom). For this analysis, we generated 100 random outgroup replicates and ran each under a MP heu- Bayesian analysesÐModels of molecular evolution were chosen for each ristic search with ®ve random addition replicates and TBR branchswapping. data set and each gene region using Modeltest version 3.05b (Posada and Best trees from each search were saved to a tree ®le, and outgroup number, Crandall, 1998). For any data set containing more than one gene region (i.e., identity, and root placement was recorded for all trees. all analyses with the exception of the nuclear data set), we ran a mixed-model analysis, allowing each gene region to evolve under its own best-®t model. RESULTS All Bayesian analyses were performed assuming default prior parameter distributions set in MrBayes v.3.04b. Posterior probabilities of trees were ap- Taxon sampling and missing dataÐWhile our combined proximated using the MCMC algorithm with four incrementally heated chains data set includes 38 taxa, not all taxa were sequenced for all (T ϭ 0.2) for 10 000 000 generations and sampling trees every 1000 genera- ®ve gene regions (Table 2; Appendix). We were unable to tions. To estimate appropriate burn-ins, posterior parameter distributions were amplify the phyC region from two of our outgroup taxa, Cer- viewed using Tracer v.1.01 (Rambaut and Drummond, 2003). Discarding the aria fruticulosa and Portulacaria afra.InthetrnK/matK data ®rst 100 trees (100 000 generations) was adequate in ensuring stationarity. set, 36 of 38 taxa are sequenced for the matK gene and ¯anking 3Ј trnK intron, while a subset of those taxa are also se- Hypothesis testing and sensitivity analysesÐTo test for alternative posi- quenced for the ¯anking 5Ј trnK intron. Also, two Pereskia tions of the root of Cactaceae, two parametric bootstrap tests were performed species (Pereskia bahiensis and Pereskia diaz-romeroana) (Goldman et al., 2000). Our ®rst test constrained Pereskia to be monophyletic, were only sequenced for the trnK/matK region; this region and in this case, our outgroups rooted along the branch to the Opuntioideae. provided substantial information regarding the placement of In our second test, we constrained the position of the root along the branch these taxa within two strongly supported Pereskia clades. Fi- to the Andean and southern South American Pereskia clades (see Discussion, nally, two taxa in our data set are chimeric: Cereus comprises Pereskia paraphyly and ovary position). This scenario is consistent with one ®rst proposed by Berger (1926), in which his Eupereskia, a subgenus corre- C. alacriportanus (trnK/matK) and C. fernambucensis (all oth- sponding roughly to our Andean and southern South American Pereskia er regions), and Pereskiopsis A comprises P. diguetii (trnK/ clades, is sister to all other cacti. matK) and P. aquosa (all other regions). In each case, a ML heuristic search was performed on the original combined data set with tree topology T0 constrained to the alternative hypothesis, Data partition combinabilityÐThe ILD test rejected the ␦ϭ Ϫ and the test statistic LML LT was calculated. Using SeqGen v.1.2.7 null hypothesis that our two data partitions were derived from (Rambaut and Grassly, 1997), we simulated 100 replicate data sets using the the same data pool (P ϭ 0.045). Visual inspection of trees

best-®t model of evolution for the original data set and tree topology T0 with from analyses of the individual partitions revealed two distinct 1180 AMERICAN JOURNAL OF BOTANY [Vol. 92

areas of incongruence (Fig. 1), and removal of either the rep- resentatives of the Anacampseroteae (ϭ Grahamia bracteata, Grahamia coahuilensis, and Anacampseros telephiastrum) outgroup taxa or the clade comprising P. zinnii¯ora, P. portulacifolia, P. quisqueyana, and P. marcanoi from the ILD analysis resulted in P Ͼ 0.05. As accurate resolution of either tree region is not the primary purpose of this study, we did not investigate partition congruence issues further, and pooled all markers together for a combined analysis. To be certain that incongruence in one part of the tree was not affecting our analysis in other parts of the tree, we removed the aforementioned taxa from the data set and re-ran our MP analyses. The resulting trees for the remaining taxa were topologically identical to those recovered from analyses of the full data set.

Separate partition analysesÐThe different gene regions sequenced vary widely in the amount of phylogenetic information they contain, with the nuclear marker phyC offering mark- edly greater resolution at the deeper nodes within Cactaceae (Table 2; Fig. 1).

cpDNA/mtDNA markersÐThe 1000 MP heuristic searches resulted in 26 shortest trees with length 1173 and CI ϭ 0.83 (CI excluding autapomorphies ϭ 0.69). The 100 heuristic searches using maximum likelihood and the HKY85 model of evolution resulted in a single tree with score ϪlnL ϭ 14937.27192 (Fig. 2). Bootstrap support was weak for deeper nodes within the tree using either MP or ML (unresolved in Ͻ Fig. 1. Majority-rule consensus trees built from most parsimonious trees Fig.1, support values 50%). There is, however, strong MP, found in 1000 heuristic parsimony searches of each of the two data partitions. ML, and Bayesian support for Cactaceae, Cactoideae, and Four taxa were pruned from the cpDNA/mtDNA tree that were not sequenced Opuntioideae, as well as several clades of Pereskia that cor- for phyC. Numbers along branches represent support values for clades. Num- respond well with Leuenberger's (1986) species groups (Fig. bers above branches are (MP bootstraps/ML bootstraps); numbers below 1). Pereskia guamacho and P. aurei¯ora are sister to one an- branches are Bayesian posterior probabilities. Branches with less than 50% bootstrap support values are collapsed. For full taxon names please refer to other, as indicated by Leuenberger. A group of ®ve Pereskia the Appendix. species (P. grandifolia, P. sacharosa, P. nemorosa, P. stenantha, and P. bahiensis), distributed in the cerrhado, caatinga,

Fig. 2. MLE phylogram trees from separate data partitions and combined analysis. For full taxon names please refer to the Appendix. July 2005] EDWARDS ET AL.ÐBASAL CACTUS PHYLOGENY 1181

Fig. 3. Bayesian consensus tree for combined analysis. Numbers along branches represent support values for clades; numbers above branches are (MP bootstraps/ML bootstraps). Numbers below branches are Bayesian posterior probabilities. Bootstrap values less than 50% are represented by dashes. and chaco woodland ecosystems of southern South America, The Caribbean island endemics are subtended by P. bleo, is well supported and is hereafter referred to as the SSA (ϭ which has been described as a riparian species, and is distrib- southern South America) clade (Fig. 3). There is also strong uted in forests of Panama and Colombia. We will refer to support for uniting the four Pereskia species restricted to the (Pereskia bleo (P. quisqueyana, P. marcanoi, P. portulacifol- Caribbean islands of Hispaniola and Cuba. The cpDNA/ ia, P. zinnii¯ora)) as the Caribbean clade (Fig. 3). Pereskia mtDNA data set strongly suggests that the species endemic to aculeata, a geographically widespread and ecologically di- Hispaniola (P. quisqueyana, P. portulacifolia, and P. marca- verse species with a semi-scandent habit, is placed as sister to noi) are not monophyletic; instead, P. zinnii¯ora, an endemic a group of three morphologically similar species (P. horrida, to the island of Cuba, is the sister species of P. portulacifolia. P. diaz-romeroana, and P. weberiana) that live in the dry 1182 AMERICAN JOURNAL OF BOTANY [Vol. 92

Inter-Andean valley regions of Peru and Bolivia. Leuenberger a single ML tree with ϪlnL ϭ 20373.64. MP, ML, and Bayes- united the three Andean species but could not place P. acu- ian analyses all resulted in an identical tree topology, consist- leata. We will refer to (P. aculeata (P. horrida, P. diaz-rom- ing of 3 major clades: the Northern Pereskia clade, the Opun- eroana, and P. weberiana)) as the Andean clade (Fig. 3). tioideae, and Maihuenia ϩ Cactoideae. Pereskia is paraphyletic, with the Northern Pereskia clade placed as the sister of Phytochrome CÐOur phyC fragments correspond to the all remaining cacti. The Andean and SSA Pereskia lineages exon 1 region of Mathews and Donoghue (2000). Our MP in turn form a weakly supported clade that appears to share a heuristic search recovered 48 trees with length 665 and CI ϭ sister relationship with the ``core cacti'' (Fig. 3), consisting of 0.75 (CI excluding autapomorphies ϭ 0.63). ML recovered a the opuntioid, cactoid, and Maihuenia lineages. We will refer single best tree with ϪlnL ϭ 5428.33231 (Fig. 2). MP, ML, to ((Andean, SSA)(Opuntioideae (Maihuenia, Cactoideae))) as and Bayesian analyses recover all of the aforementioned clades the caulocactus clade (caulo ϭ stem, referring to their spe- supported by the chloroplast±mitochondrial data set. Addition- cialized stems, see Discussion, early events in the evolution ally, phyC resolves a basal split in the Cactaceae between one of stem-based photosynthesis). clade of Pereskia and all other cacti. The basal Pereskia clade For the most part, the phyC and combined trees are in agree- comprises P. lychnidi¯ora, P. guamacho ϩ P. aurei¯ora, and ment, with the major exception being the formation of the core the Caribbean clade. This basal split is strongly supported by cacti lineage in the combined analysis. Bayesian support for ML bootstraps and Bayesian posterior probabilities, but less the core cacti is strong, ML bootstrap support is less so, and so by parsimony (Fig. 1). This clade comprises all of the Per- the MP bootstrap value is less than 50%. eskia species with distributions centered loosely around the Caribbean basin. We will refer to this newly identi®ed Per- Testing for alternative rootingsÐOur conclusion of a par- eskia lineage as the Northern Pereskia clade (Fig. 3). aphyletic Pereskia is dependent upon the assumption that the The Andean and SSA Pereskia clades are united with the outgroup taxa are attaching along the correct branch in Cac- opuntioid, cactoid, and Maihuenia lineages. Maihuenia and taceae. Problems with incorrect outgroup attachment become Cactoideae, in turn, are placed in a sister relationship. Within more likely when chosen outgroups are distantly related to the the Cactoideae, Blossfeldia liliputana is placed as sister to the ingroup (Sanderson and Doyle, 2001; Graham et al., 2002; other cactoids included in this study. This supports the con- Soltis and Soltis, 2004). While we feel that our sample of troversial ®nding of Nyffeler (2002) that B. liliputana, a di- outgroup taxa probably includes the closest living relatives of minutive, globular cactus of northern Argentina and Bolivia, the cacti (Hershkovitz and Zimmer, 1997; Applequist and Wal- is the basal member of Cactoideae. While the sampling of lace, 2001; R. Nyffeler, unpublished data), there is consider- Cactoideae in this study is clearly too limited to hypothesize able sequence divergence among the outgroups, as evidenced major cactoid relationships, it is encouraging that phyC, a nu- by the long branch lengths in Fig. 2. As described earlier, we clear gene, is so far in agreement with the results of Nyffeler investigated this potential problem in two ways. First, we used (2002), which was based on only chloroplast markers. parametric bootstrapping to compare two alternative rooting Two areas of incongruence between our nuclear and cp- hypotheses. Second, we conducted an outgroup jackkni®ng ex- DNA/mtDNA data sets are worth noting (Fig. 1). The ®rst periment to see how different combinations of available out- concerns the arrangement of taxa in the Portulaca/Talinum/ groups might alter the placement of the root. Anacampseroteae outgroups, with the chloroplast sequences Results from our parametric bootstrapping tests strongly re- suggesting a Cactaceae-Anacampseroteae sister relationship, ject a monophyletic Pereskia (original data set, P Ͻ 0.01, re- and phyC suggesting that Portulaca is the nearest living rel- duced data set, P Ͻ 0.01), as well as a rooting along the ative of Cactaceae. Because Portulacaria and Ceraria were branch to the Andean ϩ SSA clade (original, P Ͻ 0.01, re- not included in our phyC analysis, we thought this may be a duced, P Ͻ 0.01). In addition, results from our outgroup jack- result of differential outgroup taxon sampling. Re-analyzing kni®ng experiment provide some evidence that rooting along our cpDNA/mtDNA data set without Portulacaria and Cer- the Northern Pereskia clade is not an artifact of outgroup tax- aria, however, did not change the branching order of these on sampling. Of 100 random replicate outgroup samplings, six outgroups. outgroup combinations resulted in alternative rootings along The second incongruence lies within the well-supported either the opuntioid or cactoid branches. In each of these six Pereskia clade restricted to the Caribbean islands of Cuba and cases, three or fewer outgroup taxa were present, and Ceraria Hispaniola. Our chloroplast data places the Cuban endemic P. fruticulosa and/or Portulacaria afra were always included. zinnii¯ora within a clade comprising the Hispaniolan endemics These two taxa are the most distantly related outgroups in our P. quisqueyana, P. portulacifolia, and P. marcanoi. The nu- analysis (Hershkovitz and Zimmer, 1997; Applequist and Wal- clear data, on the other hand, ®rmly resolve the Hispaniolan lace, 2001; R. Nyffeler, unpublished data). Within the Talin- endemics as a monophyletic group sharing a sister relationship um/Portulaca/Anacampseroteae lineages, nearly all taxa root- with the Cuban P. zinnii¯ora. Relationships among these four ed along the branch to the Northern Pereskia clade, whether taxa deserve additional attention. The three Hispaniolan spe- they were alone or present in any other combination. The ex- cies currently exist as a handful of small, rather isolated pop- ception was Anacampseros telephiastrum, which rooted to the ulations, though it is not impossible to imagine gene ¯ow oc- Northern Pereskia clade when it was the only outgroup, but curring between them. rooted along the opuntioid or cactoid branches when it was present with C. fruticulosa or P. afra. Combined analysesÐCombining all gene regions into one data set yielded a well-supported hypothesis of basal cactus DISCUSSION phylogeny (Figs. 2, 3). The 1000 heuristic MP searches resulted in 12 trees of length 1846 and CI ϭ 0.80 (CI excluding Pereskia paraphyly and ovary positionÐThe tree shown in autapomorphies ϭ 0.66). The 100 heuristic ML searches found Fig. 3, in which Pereskia is paraphyletic, is strongly supported July 2005] EDWARDS ET AL.ÐBASAL CACTUS PHYLOGENY 1183 by our analyses. However, most of the basal structure of our (usually microscopic and/or ephemeral) is borne subtending tree is contributed by the phyC gene, so it may be more ap- the areole, while in some Pereskia species, leaves are also propriate at this stage to consider this to be a gene tree rather borne directly on the areoles (Gibson and Nobel, 1986). This than the cactus species tree. While the implications of Pereskia is reminiscent of the short-shoot systems of many desert paraphyly should be considered at this stage, we would like plants, which are utilized to produce quick ¯ushes of leaves to see our results con®rmed with additional genes before em- without ®rst investing in new stem tissue. Because some Per- barking on a new nomenclature for Cactaceae. eskia species and all core cacti lack this trait, a reasonable The ®rst hypothesis of Pereskia paraphyly dates back to scenario in the transition to lea¯essness would be ®rst to lose Berger (1926, ``Schema 1''), who split the genus into Euper- the ability to ¯ush leaves from the areole. While this may still eskia and Rhodocactus and united Rhodocactus with Opun- be the case (as likely in the Cylindropuntieae; see Wallace and tioideae, Cactoideae, and Maihuenia. His decision to split Per- Dickie, 2002), extant Pereskia species provide us with no clear eskia in this way was based on his interpretation of the Per- record of it. Areole leaf production appears to have a compli- eskia ovary, with Eupereskia having a distinctly superior ova- cated history, with potentially two losses (in the caulocactus ry with basal placentation and Rhodocactus having parietal and Caribbean clades) and then two subsequent gains (within placentation and a more developed receptacle surrounding the the Caribbean and SSA clades), or three gains (in the North- ovary. From later work, it appears that Berger had access to ern, Caribbean, and SSA clades) and one loss (within the Ca- ®ve Pereskia species (P. sacharosa, P. aculeata, ribbean clade) (Fig. 4B). P.grandifolia, P. bleo, and P. lychnidi¯ora) (Berger, 1929). In contrast, there are two stem characters that clearly unite His two Pereskia genera correspond with the Northern (ϭ the Andean and SSA Pereskia clades with the core cacti. Most Rhodocactus) and Andean ϩ SSA (ϭ Eupereskia) lineages, members of the caulocactus clade produce stomata on their with the exception of P. grandifolia, which he united with the stem epidermis and have the delayed bark formation charac- Northern species. However, his idea of family level relation- teristic of the stem photosynthesizing cacti (Fig. 4C, D). A ships (Eupereskia (Rhodocactus, Cactoideae, Opuntioideae, notable exception is Maihuenia. While Mauseth (1999) dis- Maihuenia)) was rejected in our parametric bootstrap tests. cussed stomata on fragmented epidermis in the areolar pits of Subsequent Pereskia classi®cations followed on this theme, both Maihuenia species, they lacked stomata in all other stem though they have proposed slight rearrangements of taxa and regions, so we interpret Maihuenia stems here to be essentially have differed in the assignment of genus or subgenus rank astomatous (Leuenberger, 1997, also describes them as lacking (Backeberg and Knuth, 1936; Backeberg, 1958). All these stem stomata). The placement of Maihuenia in our trees im- schemes have been based primarily on differences in ovary plies that both stem stomata and delayed bark formation have position because it was generally agreed that the variability of been lost in this lineage. Maihuenia today consists of two the Pereskia gynoecium held the key to understanding the evo- highly specialized cushion plant species living in cold, dry lution of the ``sunken ovary'' of the core cacti. In spite of this regions of Patagonia and south-central Chile (Leuenberger, interest, however, the Pereskia gynoecium did not receive 1997; Mauseth, 1999). It is not immediately clear why these careful study until much later (Boke, 1963, 1964, 1966, 1968; traits have been lost, though their cushion plant growth habit Leins and Schwitalla, 1988), and earlier misinterpretations are (often forming very dense stands of stems with leaves crowded likely responsible for the confusing history of Pereskia clas- at tips) seems unlikely to promote the use of the stem as a si®cation. Leuenberger (1986) stressed the complex nature of photosynthetic organ. the ovary in Pereskia, noting that ``the transitional nature of All members of the Northern Pereskia clade lack stem sto- ovary and ovule position, subject to distortions during ontog- mata and delayed bark formation, though stomata are very eny of ¯ower and fruit, makes it an unreliable character'' (p. rarely found on the stems of P. portulacifolia and P. guama- 51). Pereskia sacharosa, for example, has been described as cho (E. Edwards, personal observation). It is currently impos- having ``fully superior'' and ``half inferior'' ovaries, with both sible to ascertain whether these traits were lost in the Northern ``basal'' and ``basal-parietal'' placentation, depending on the clade or gained in the caulocactus clade because we are still author. Leuenberger concluded that most Pereskia species uncertain as to the true sister taxa to Cactaceae, and there is could be considered half superior or half inferior, with truly little mention of these traits in the Portulacaceae literature inferior ovaries found only in the Caribbean clade (Fig. 4A). (Carolin, 1987, 1993; Eggli and Ford-Werntz, 2001; but see It appears that the evolutionary transition to the ``cactus-type'' Hershkovitz, 1993, for comment on stem stomata in Talinum). inferior ovary that botanists had long hoped to ®nd in Pereskia Nevertheless, we can say that stem stomata and delayed bark does exist, but as an independent event within the Caribbean formation were present in the cactus lineage prior to the di- clade. Careful comparative studies within the Northern Per- vergence of opuntioid and cactoid lineages and that the evo- eskia clade might be helpful in understanding this transition, lution of these two characters appears to have been correlated. and it would also be useful to assess the degree of similarity It is worth considering how the presence of these traits of the Caribbean ovary to the inferior ovary of the core cacti. might in¯uence the functioning of stems and leaves in Per- eskia. Several greenhouse studies have investigated the pho- Early events in the evolution of stem-based photosynthe- tosynthetic behavior of a variety of Pereskia species (Rayder sisÐWhile there are no obvious reproductive characters unit- and Ting, 1981; Nobel and Hartsock, 1986, 1987; Martin and ing the Andean and SSA Pereskia clades with the major cactus Wallace, 2000) and minimal C3 stem photosynthesis has been lineages, our phylogenetic hypothesis provides some insight recorded only in P. horrida (Martin and Wallace, 2000). It is into early events in the development of the stem as the primary dif®cult to extrapolate these ®ndings to what might happen in photosynthetic organ of the cacti. Several traits from Table 1 a natural environment, however, and ecological observations could potentially be important in the evolutionary transition to of wild Pereskia species are currently quite limited in scope. a functionally lea¯ess state. For instance, leaves occur in two Projects characterizing the ecological physiology of the dif- places in cacti. In all cacti, during new stem growth a leaf ferent Pereskia lineages are currently underway (Edwards et 1184 AMERICAN JOURNAL OF BOTANY [Vol. 92

Fig. 4. Reconstruction of ancestral character states using unordered parsimony. A. Ovary position (white ϭ superior, gray ϭ half inferior, black ϭ inferior). B. Areole leaf production (white ϭ no leaves on areoles, black ϭ leaves on areoles). C. Delayed bark formation (white ϭ precocious bark formation, black ϭ delayed). D. Stem stomata (white ϭ no stem stomata, black ϭ stem stomata). In B, C, and D, outgroups are coded as unknown. al., 2004) and should be helpful in discerning any functional cactus stem. These include a hypodermal layer comprised of differences associated with these anatomical changes. collenchymatous cells with thickened walls, substantial inter-

In the meantime, however, it seems that further modi®ca- cellular air space to aid in CO2 diffusion, and the arrangement tions of the cortical tissue are needed to make much photo- of cortical chlorenchyma tissue into ``spongy mesophyll'' and synthetic use of the stem. In a survey of 28 cacti, including ``palisade'' layers that appear similar to those in photosyn- seven Pereskia species representing the Northern, Andean, and thetic leaves. While their sampling did not include any opun- SSA clades, Sajeva and Mauseth (1991) emphasized what they tioids, other researchers have found similar cortical structure considered to be key developments in the evolution of the in Opuntioideae (Nobel, 1988; North et al., 1995). No Per- July 2005] EDWARDS ET AL.ÐBASAL CACTUS PHYLOGENY 1185

Fig. 5. Geographic distribution of basal cacti. Species ranges correspond to shaded area, except for the geographically widespread taxa: Pereskia aculeata (a circle), the Opuntioideae (a triangle), and all Cactoideae minus Blossfeldia (a triangle). eskia species included in the Sajeva and Mauseth (1991) study to function as an ef®cient photosynthetic organ (cf. ``devel- had their cortical tissue differentiated into such layers, and opmental enablers'' in Donoghue, 2005). This order of events further, all of their stems had relatively small volumes of in- may not be unique to cacti: in a recent survey of noncactus tercellular air space. It appears from our phylogenetic results plants with stem-based photosynthesis, Mauseth (2004) found that the evolution of stem stomata and delayed bark formation that the majority of investigated species with delayed bark preceded the key cortex modi®cations that allowed the stem formation, potentially photosynthetic stems, and photosyn- 1186 AMERICAN JOURNAL OF BOTANY [Vol. 92 thetic leaves did not have a specialized cortex, while nearly ConclusionsÐPereskia, long thought to represent the ``an- all of the aphyllous species with stems as their primary pho- cestral'' cactus, appears to be paraphyletic, with the Andean tosynthetic organ also had a well-developed palisade cell layer and SSA Pereskia clades more closely related to the core cacti. in their outer cortical tissue. This result allows us to make several new inferences regarding early cactus evolution. First, it appears that the specialized ``sunken'' inferior ovary of the cacti has evolved twice inde- Historical biogeography of the cactiÐThere have been sev- pendently, once in the core cacti, and once in the Caribbean eral hypotheses regarding the geographical origin of the cacti, Pereskia clade. Second, most members of the two Pereskia all largely based on where the presumably basal members of clades that are united with the core cacti delay bark formation Cactaceae currently reside. Buxbaum (1969) cited both the Ca- and produce stem stomata. As these Pereskia clades lack other ribbean and central South America as likely areas, due to the stem specializations found in the core cacti (e.g., differentia- presence of Pereskia and of opuntioid and cactoid lineages tion of cortical tissue into specialized cell types, large inter- that he considered ``ancestral.'' Leuenberger (1986) concluded cellular air spaces, well developed hypodermis), we infer that that northwestern South America was a more reasonable lo- the expression of stem stomata and delayed bark formation cation, primarily because he imagined a late Cretaceous origin were among the ®rst changes associated with the transition to of the group and because centering its origin as far away from stem-based photosynthesis. Finally, our phylogenetic hypoth- Africa as possible might explain the poor representation of esis implies an initial north±south geographic split during early cacti there. Our data set, as well as other recent molecular cactus evolution and that both Opuntioideae and Cactoideae phylogenetic studies (Hershkovitz and Zimmer, 1997; Nyffel- originated in southern or west central South America. er, 2002), suggest that the cacti are not that old, however, be- While these results are promising, they also point to areas cause sequence divergence between the major cactus clades is that need work. First and foremost, we would like to see our limited. More recently, Wallace and Gibson (2002) hypothe- results tested with additional genes before reclassifying the sized a central Andean origin for the family. They assumed Cactaceae. Clearly, the true position of the cacti within the that the Andean Pereskia, certain Opuntioideae, and the cac- portulacaceous alliance needs to be settled, as this is critical toid Calymmanthium are basal cactus lineages, and all reside to interpretations of early cactus character evolution. Perhaps in Peru, Bolivia, and northern Argentina. most important, however, is the need for careful study of Per- Our ability to infer the geographic origin of Cactaceae is eskia ecology. The specialized physiology and water use strat- limited by insuf®cient knowledge of its closest relatives, but egies of the core cacti have been well characterized (Nobel, within the cacti there is a striking geographic structure to our 1988), while virtually no studies have been conducted on Per- hypothesis of basal relationships (Fig. 5). The basal split in eskia in the wild (but see Diaz, 1984; Diaz and Medina, 1984). Cactaceae suggests an initial separation within the cacti into a The cactus life form represents one of the more extreme cases primarily northern and a primarily southern clade. The North- of ecological and morphological specialization in plants, and ern Pereskia clade is comprised of species scattered in Central knowledge that (and precisely how) Pereskia is paraphyletic and northern South America, Cuba, and Hispaniola, with the provides a unique opportunity to examine the earliest steps in exception of P. aurei¯ora of eastern Brazil. Within the cau- its evolution. Discovering more about how members of the locactus clade, Maihuenia is restricted to central/southern different Pereskia lineages function in their environments Chile and Argentina, the Andean Pereskia clade (with the ex- would complement what we already know about their anatomy ception of P. aculeata, the only geographically widespread and morphology, allowing for an integrated approach to un- Pereskia species) is found only in Peru and Bolivia, and the derstanding this remarkable evolutionary transition. SSA clade is scattered throughout drier forest regions of east- LITERATURE CITED ern Brazil, Paraguay, Uruguay, Argentina, and Bolivia. Bloss- feldia is con®rmed here as a basal cactoid lineage, and grows APPLEQUIST,W.L.,AND R. S. WALLACE. 2001. Phylogeny of the portula- only in northern Argentina and Bolivia. While our sampling caceous cohort based on ndhF sequence data. Systematic Botany 26: 406± and resolution are insuf®cient within Opuntioideae to make 419. BACKEBERG, C. 1958. Die Cactaceae. Handbuch der Kakteenkunde. G. Fi- inferences about the geographic distribution of its basal mem- scher, Jena, Germany. bers, recent work of others has placed its earliest diverging BACKEBERG, C., AND F. M. KNUTH. 1936. Kaktus-ABC. Nordisk Forlag, Co- lineages in Chile and Argentina (Grif®th, 2004). penhagen, Denmark. Our results imply that both Opuntioideae and Cactoideae BAILEY, I. W. 1962. The xylem of Pereskia sacharosa and Pereskia aculeata. originated in the southern half of South America, possibly in Journal of the Arnold Arboretum 43: 376±388. BAILEY, I. 1963a. The xylem of Pereskia grandifolia and Pereskia bleo. the central Andean region of Peru, Bolivia, and northern Ar- Journal of the Arnold Arboretum 44: 222±231. gentina. Andean orogeny has been viewed as an important BAILEY, I. 1963b. The xylem of pereskias from Peru and Bolivia. Journal of cause of diversi®cation for other plant lineages (Raven and the Arnold Arboretum 44: 127±137. Axelrod, 1974; Burnham and Graham, 1999). Early uplift in BAILEY, I. 1963c. The xylem of pereskias from southern Mexico and Central the central Andean region (ϳ25±20 mya) is hypothesized to America. Journal of the Arnold Arboretum 44: 211±221. BAILEY, I. 1963d. The xylem of Pereskia colombiana, Pereskia guamacho, have occurred under a ¯uctuating arid/semi-arid climate re- Pereskia cubensis, and Pereskia portulacifolia. Journal of the Arnold gime (Hartley, 2003), which presents a likely scenario for early Arboretum 44: 390±401. cactus diversi®cation. The placement of Cacteae, an exclu- BARTHLOTT,W.,AND D. R. HUNT. 1993. Cactaceae. In K. Kubitzki, J. G. sively North American lineage with its center of diversity in Rohwer, and V. Bittrich [eds.], The families and genera of vascular Mexico, as being among the earliest diverging lineages of Cac- plants, 161±197. Springer Verlag, Berlin, Germany. BERGER, A. 1926. Die entwicklungslinien der Kakteen. G. Fisher, Jena, Ger- toideae (represented here with Echinocactus, but con®rmed many. with broader sampling in Nyffeler, 2002), suggests relatively BERGER, A. 1929. Illustrierte handbucher sukkulenter p¯anzen: Kakteen. Ul- early movements out of this region and across the continent. mer, Stuttgart, Germany. July 2005] EDWARDS ET AL.ÐBASAL CACTUS PHYLOGENY 1187

BOKE, N. H. 1954. Organogenesis of the vegetative shoot in Pereskia. Amer- LEUENBERGER, B. E. 1997. MaihueniaÐmonograph of a Patagonian genus ican Journal of Botany 41: 619±637. of Cactaceae. Botanische Jahrbucher 119: 1±92. BOKE, N. H. 1963. Anatomy and development of the ¯ower and fruit of MADDISON,D.R.,AND W. P. M ADDISON. 2003. MacClade v. 4.06. Sinauer Pereskia pititache. American Journal of Botany 50: 843±858. Associates, Sunderland, Massachusetts, USA. BOKE, N. H. 1964. The cactus gynoecium: a new interpretation. American MARTIN,C.E.,AND R. S. WALLACE. 2000. Photosynthetic pathway variation Journal of Botany 51: 598±610. in leafy members of two subfamilies of the Cactaceae. International BOKE, N. H. 1966. Ontogeny and structure of the ¯ower and fruit of Pereskia Journal of Plant Sciences 161: 639±650. aculeata. American Journal of Botany 53: 534±542. MATHEWS, S., AND M. J. DONOGHUE. 1999. The root of angiosperm phylog- BOKE, N. H. 1968. Structure and development of the ¯ower and fruit of eny inferred from duplicate phytochrome genes. Science 286: 947±950. Pereskia diaz-romeroana. American Journal of Botany 55: 1254±1260. MATHEWS, S., AND M. J. DONOGHUE. 2000. Basal angiosperm phylogeny BOUSQUET, J., S. H. STRAUSS, AND P. L I. 1992. Complete congruence be- inferred from duplicate phytochromes A and C. International Journal of tween morphological and rbcL-based molecular phylogenies in birches Plant Sciences 161: S41±S55. and related species (Betulaceae). Molecular Biology and Evolution 9: MAUSETH, J. D. 1999. Anatomical adaptations to xeric conditions in Mai- 1076±1088. huenia (Cactaceae), a relictual, leaf-bearing cactus. Journal of Plant Re- BURNHAM,R.J.,AND A. GRAHAM. 1999. The history of neotropical vege- search 112: 307±315. tation: new developments and status. Annals of the Missouri Botanical MAUSETH, J. D. 2004. The structure of photosynthetic succulent stems in plants Garden 86: 546±589. other than cacti. International Journal of Plant Sciences 165: 1±9. BUXBAUM, F. 1969. Die entwicklungswege der Kakteen in Sudamerika. In E. MAUSETH,J.D.,AND J. V. LANDRUM. 1997. Relictual vegetative anatomical J. Fittkau, J. Illies, H. Klinge, G. H. Schwabe, and H. Sioli [eds.], Bio- characters in Cactaceae: the genus Pereskia. Journal of Plant Research geogaphy and ecology in South America, 583±623. W. Junk, The Hague, 110: 55±64. Netherlands. MAUSETH, J. D., Y. UOZUMI,B.J.PLEMONS, AND J. V. LANDRUM. 1995. CAROLIN, R. 1987. A review of the family Portulacaceae. Australian Journal Structural and systematic study of an unusual tracheid type in cacti. Jour- of Botany 35: 383±412. nal of Plant Research 108: 517±526. CAROLIN, R. 1993. Portulacaceae. In K. Kubitzki [ed.], Families and genera NOBEL, P. S. 1988. Environmental biology of agaves and cacti. Cambridge of vascular plants, 544±555. Springer-Verlag,Heidelberg, Germany. University Press, New York, New York, USA.

DIAZ, M. 1984. Estudios ®soecologicos de 4 especies de cactaceas en con- NOBEL,P.S.,AND T. L. HARTSOCK. 1986. Leaf and stem CO2 uptake in the diciones naturales. Masters thesis, Instituto Venezolano de Investiga- 3 subfamilies of the Cactaceae. Plant Physiology 80: 913±917. ciones Cienti®cas, Caracas, Venezuela. NOBEL,P.S.,AND T. L. HARTSTOCK. 1987. Drought-induced shifts in daily

DIAZ, M., AND E. MEDINA. 1984. Actividad CAM de cactaceaes en condi- CO2 uptake patterns for leafy cacti. Physiologia Plantarum 70: 114±118. ciones naturales. In E. Medina [ed.], Eco-Fisiologia de Plantas CAM. NORTH, G. B., T. LIN MOORE, AND P. S. NOBEL. 1995. Cladode development

Centro Internacional de Ecologia Tropical,Caracas, Venezuela. for Opuntia ®cus-indica (Cactaceae) under current and doubled CO2 con- DONOGHUE, M. J. 2005. Key innovations, convergence, and success: mac- centrations. American Journal of Botany 82: 159±166. roevolutionary lessons from plant phylogeny. Paleobiology 31: 77±93. NYFFELER, R. 2002. Phylogenetic relationships in the cactus family (Cacta- DUMINIL, J., M. H. PEMONGE, AND R. J. PETIT. 2002. A set of 35 consensus ceae) based on evidence from trnK/matK and trnL-trnF sequences. primer pairs amplifying genes and introns of plant mitochondrial DNA. American Journal of Botany 89: 312±326. Molecular Ecology Notes 2: 428±430. POSADA, D., AND K. A. CRANDALL. 1998. MODELTEST: testing the model EDWARDS, E. J., M. DIAZ, AND M. J. DONOGHUE. 2004. Water relations of of DNA substitution. Bioinformatics 14: 817±818. Pereskia guamacho, a cactus with leaves. Botany 2004: Annual meeting RAMBAUT, A., AND A. J. DRUMMOND. 2003. Tracer v. 1.0.1, Available from of the Botany Society of America, Snowbird, Utah USA, available website http://evolve.zoo.ox.ac.uk/. at website http://www.2004.botanyconference.org/engine/search/index. RAMBAUT, A., AND N. C. GRASSLY. 1997. Seq-Gen: an application for the php?funcϭdetail&aidϭ663. Monte Carlo simulation of DNA sequence evolution along phylogenetic EGGLI, U., AND D. FORD-WERNTZ. 2001. Portulacaceae. In U. Eggli [ed.], trees. Bioinformatics 13: 235±238. Illustrated handbook of succulent plants: dicotyledons, 370±432. Spring- RAUH, W. 1979. Kakteen an ihren standorten. Parey, Berlin, Germany. er-Verlag, Berlin, Germany. RAVEN,P.H.,AND D. I. AXELROD. 1974. Angiosperm biogeography and past FARRIS, J. S., M. KAÈ LLERSJOÈ ,A.G.KLUGE, AND C. BULT. 1994. Testing continental movements. Annals of the Missouri Botanical Garden 61: signi®cance of incongruence. Cladistics 10: 315±319. 539±673. GIBSON,A.C.,AND P. S. NOBEL. 1986. The cactus primer. Harvard Univer- RAYDER, L., AND I. P. TING. 1981. Carbon metabolism in 2 species of Per- sity Press, Cambridge, Massachusetts, USA. eskia (Cactaceae). Plant Physiology 68: 139±142. GOLDMAN, N., J. P. ANDERSON, AND A. G. RODGRIGO. 2000. Likelihood- SAJEVA, M., AND J. D. MAUSETH. 1991. Leaf-like structure in the photosyn- based tests of topologies in phylogenetics. Systematic Biology 49: 652± thetic, succulent stems of cacti. Annals of Botany 68: 405±411. 670. SANDERSON, M. J., AND J. A. DOYLE. 2001. Sources of error and con®dence GRAHAM, S. W., R. G. OLMSTEAD, AND S. C. H. BARRETT. 2002. Rooting in estimating the age of angiosperms from rbcL and 18s rDNA data. phylogenetic trees with distant outgroups: a case study from the com- American Journal of Botany 88: 1499±1516. melinoid monocots. Molecular Biology and Evolution 19: 1769±1781. SANG, T., D. J. CRAWFORD, AND T. F. S TUESSY. 1997. Chloroplast DNA GRIFFITH, M. P. 2004. What did the ®rst cactus look like? An attempt to phylogeny, reticulate evolution, and biogeography of Paeonia (Paeoni- reconcile the morphological and molecular evidence. Taxon 53: 493±499. aceae). American Journal of Botany 84: 1120±1136. HARTLEY, A. J. 2003. Andean uplift and climate change. Journal of the Geo- SCHLEIDEN, M. J. 1845. BeitraÈge zur anatomi der Kacteen. MeÂmoires de l' logical Society, London 160: 7±10. AcadeÂmie ImpeÂeriale des Sciences de St-PeÂtersbourg 4: 335±380. HERSHKOVITZ, M. A. 1993. Revised circumscriptions and subgeneric taxon- SCHUMANN, K. 1898. Gesamtbeschreibung der Kakteen. Neumann, Neu- omies of Calandrinia and Montiopsis (Portulacaceae) with notes on phy- damm, Germany. logeny of the portulacaceous alliance. Annals of the Missouri Botanical SOLTIS,D.E.,AND P. S. SOLTIS. 2004. Amborella not a ``basal angiosperm''? Garden 80: 333±365. Not so fast. American Journal of Botany 91: 997±1001. HERSHKOVITZ,M.A.,AND E. A. ZIMMER. 1997. On the evolutionary origins SWOFFORD, D. L. 2002. PAUP*: phylogenetic analysis using parsimony (*and of the cacti. Taxon 46: 217±232. other methods), version 4. Sinauer, Sunderland, Massachusetts, USA. HUELSENBECK,J.P.,AND F. R ONQUIST. 2001. MRBAYES: Bayesian infer- WALLACE, R. S. 1995. Molecular systematic study of the Cactaceae: using ence of phylogeny. Bioinformatics 17: 754±755. chloroplast DNA variation to elucidate cactus phylogeny. Bradleya 13: JOHNSON,L.A.,AND D. E. SOLTIS. 1994. matK DNA sequences and phylo- 1±12. genetic reconstruction in Saxifragaceae. Systematic Botany 19: 143±156. WALLACE,R.S.,AND S. L. DICKIE. 2002. Systematic implications of chlo- LEINS,P.,AND S. SCHWITALLA. 1988. Placentation in Cactaceae. In P. Leins, roplast DNA sequence variation in the Opuntioideae. In D. R. Hunt [ed.], S. C. Tucker, and P. K. Endress [eds.], Aspects of ¯oral development. J. Studies in the Opuntioideae (Cactaceae). David Hunt, Sherborne, UK. Cramer, Berlin-Stuttgart, Germany. WALLACE,R.S.,AND A. C. GIBSON. 2002. Evolution and systematics. In P. LEUENBERGER, B. E. 1986. Pereskia (Cactaceae). Memoirs of the New York S. Nobel [ed.], Cacti: biology and uses, 1±21. University of California Botanical Garden 41: 1±141. Press, Berkeley, California, USA. 1188 AMERICAN JOURNAL OF BOTANY [Vol. 92

APPENDIX. Taxa used in this study, GenBank accession numbers for the ®ve regions studied, source (wild, or if cultivated, location and accession number), and voucher information. The following abbreviations are used for herbaria and botanical gardens: YU ϭ Yale University, B ϭ Berlin Botanical Garden, ZSS ϭ Sukkulenten-Sammlung, ZuÈrich, Z ϭ ZuÈrich Botanical Garden, NTG ϭ National Tropical Garden-Kampong, Miami, JBN ϭ JardõÂn BotaÂnico Nacional, Dominican Republic, DBG ϭ Desert Botanical Garden, Phoenix, ISC ϭ University of Iowa.

Taxon; GenBank accession: phyC, rbcL, cox3, psbA-trnH, trnK/matK; Source; Voucher specimen.

Anacampseros telephiastrum D.C.; AY875311, AY875247, AY875270,Ð,Ð; Pereskia guamacho F.A.C. Weber; AY875291, AY875242, AY875254, Cult. ZSS 90 1682/10; Lavranos & Bleck s.n., South Africa, ZSS. AY875335,Ð; Wild; Edwards 15, Venezuela, YU. Pereskia guamacho F.A.C. Weber;Ð,Ð,Ð,Ð, AY015275; Cult. B 001-16- Anacampseros telephiastrum D.C.;Ð,Ð,Ð,Ð, AY875373; Cult. ZSS 90 74-70; Schwerdtfeger 13066, garden, B. 16523; Lavranos & Bleck s.n., South Africa, ZSS. Pereskia horrida (Kunth) D.C. var. horrida; AY875287, AY875224, Austrocylindropuntia subulata (Muehlenpf.) Backbg.; AY875305, AY875235, AY875258, AY875332, AY875356; Cult. B 256-01-82-30; Schwerdtfeger AY875281, AY875346, AY875364; Cult. B 153-19-74-80; Cubr 37075, gar- 13066, garden, B. den, B. Pereskia lychnidi¯ora D.C.; AY875286,Ð,Ð, AY875330, AY875358; Cult. Blossfeldia liluputana Werderm.; AY875301, AY875232, AY875282, B 003-04-78-10; Leuenberger & Schiers 2508, Guatemala, B. AY875348, AY875366; Cult. B 160-16-86-20; Leuenberger & Arroyo 3579, Pereskia lychnidi¯ora D.C.;Ð, AY875238, AY875255,Ð,Ð; Cult. DBG Argentina, B. 1990054502; Zimmerman 2603, Honduras, DBG. Brasilopuntia brasiliensis (Willd.) A. Berger; AY875304, AY875234, Pereskia marcanoi Areces; AY875288, AY875221, AY875267, AY875337,Ð; AY875278, AY875343, AY875370; Cult. B 153-76-74-80; Cubr 29521, Wild; Edwards 9, Dominican Republic, YU. 38452, garden, B. Pereskia marcanoi Areces;Ð,Ð,Ð,Ð, AY875360; Cult. B 259-02-82-30; Calymmanthium substerile Ritter; AY875314, AY875230, AY875250, Marcano & Cicero, Dominican Republic, B. AY875320, AY015291; Cult. ZSS 89 3442; Anon. 1437, garden, ZSS. Pereskia nemorosa Rojas Acosta; AY875296, AY875241, AY875276, Ceraria fruticulosa Pearson & Stephens;Ð, AY875218, AY875266,Ð, AY875334,Ð; Cult. B 305-01-80-70; Cubr 23639b, garden, B. AY875371; Cult. YU; Edwards 96, garden, YU. Pereskia nemorosa Rojas Acosta;Ð,Ð,Ð,Ð, AY875350; Cult. B 039-05- Cereus alacriportanus Pfeiff.;Ð,Ð,Ð,Ð, AY015313; Cult. ZSS 941313; Eg- 77-30; Schwerdtfeger 7085a, 15650, garden, B. gli et al. 2493, Brazil, Z. Pereskia portulacifolia (L.) D.C.; AY875315, AY875226, AY875264, AY875338,Ð; Wild; Edwards 11, Dominican Republic, YU. Cereus fernambucensis Lemaire; AY875293, AY875240, AY875272, Pereskia portulacifolia (L.) D.C.;Ð,Ð,Ð,Ð, AY875361; Cult. B 376-01-86- AY875328,Ð; Cult. B 166-88-83-10; Leuenberger et al. 3107, Brazil, B. 10; Zanoni 35204, Haiti, B. Echinocactus platyacanthus Link & Otto;Ð, AY875215, AY875257, Pereskia quisqueyana Liogier; AY875292, AY875220, AY875263, AY875327, AY015287; Cult. ZSS 92±16±86; Anon., Mexico, ZSS. AY875336,Ð; Wild; Edwards 7, Dominican Republic, YU. Echinocactus platyacanthus Link & Otto; AY875294,Ð,Ð,Ð,Ð; Cult. ZSS Pereskia quisqueyana Liogier;Ð,Ð,Ð,Ð, AY875352; Cult. B 259-05-82- 90 2985/0; LuÈthy 015, Mexico, ZSS. 33; Cicero s.n., Dominican Republic, B. Grahamia bracteata Gill.;Ð,Ð,Ð,Ð, AY015273; Cult. ZSS 94 1326; Leuen- Pereskia sacharosa Grisebach; AY875299, AY875222, AY875256, berger & Eggli 4184, Argentina, ZSS. AY875326, AY875363; Cult. B 133-10-82-30; Rente s.n., Bolivia, B. Grahamia bracteata Gill.; AY875308, AY875217, AY875273,Ð,Ð; Cult. B Pereskia stenantha Ritter; AY875295, AY875244, AY875271, AY875333,Ð; 142-32-94-10; Leuenberger & Eggli 4230b, Argentina, B. Cult. B 166-81-83-20; Leuenberger 3081, Brazil, B. Grahamia coahuilensis (S. Watson) G.D. Rowley; AY875310, AY875246, Pereskia stenantha Ritter;Ð,Ð,Ð,Ð, AY015276; Cult. ZSS 86 4200; Horst AY875280,Ð,Ð; Cult. B 262-01-94-40; Lautner L92/22, Mexico (Cult.), B. & Uebelmann 759, Brazil, ZSS. Grahamia coahuilensis (S. Watson) G.D. Rowley;Ð,Ð,Ð,Ð, AY875374; Pereskia weberiana K. Schumann; AY875313, AY875223, AY875259, Cult. ZSS 90 1259; Glas & Foster 1934, Mexico, ZSS. AY875331, AY875357; Cult. B 046-03-80-70; Schwerdtfeger 10206, gar- Maihuenia patagonica (Phil.) Britton & Rose; AY875303, AY875245, den, B. AY875277, AY875342, AY015281; Cult. B 030-30-88-10; Leuenberger & Pereskia zinnii¯ora D.C.; AY875290, AY875237, AY875262, AY875321, Arroyo 3850, Argentina, B. AY015277; Cult. ZSS 84 2526; Anon. 19537, Cuba (cult.), ZSS. Maihuenia poeppigii (Pfeiff.) K. Schum.; AY875309, AY875216, AY875269, Pereskiopsis aquosa (F.A.C. Weber) Britton & Rose; AY875300, AY875225, AY875329, AY015282; Cult. B 048-15-93-10; Leuenberger & Arroyo AY875251, AY875324,Ð; Cult. YU; Edwards 98, garden, YU. 4180, Chile, B. Pereskiopsis diguettii (F.A.C. Weber) Britton & Rose;Ð,Ð,Ð,Ð, AY015280; Opuntia dillenii (Ker-Gawl.) Haw.; AY875302, AY875233, AY875283, Cult. ZSS 94 2160; Lomeli et al. s.n., Mexico, ZSS. AY875341, AY875369; Cult. B 304-11-99-10; Greuter s.n. (4 December Pereskiopsis porteri (Brandegee ex F.A.C. Weber) Britton & Rose; 1999), Dominican Republic, B. AY875306, AY875243, AY875275, AY875344,Ð; Cult. B 169-03-84-30; Pereskia aculeata Miller; AY875312, AY875229, AY875260, AY875323,Ð; Hohmann s.n., Mexico, B. Cult. NTG; Edwards 5, garden, YU. Portulaca oleracea L.; AY875317, AY875249, AY875284,Ð,Ð; Unk.; Ap- Pereskia aculeata Miller;Ð,Ð,Ð,Ð, AY875355; Cult. B 376-02-86-20; Zan- plequist 7, Unk., ISC. oni 36108, Dominican Republic, B. Portulaca oleracea L.;Ð,Ð,Ð,Ð, AY875349; Wild; Nyffeler s.n., Switzer- Pereskia aurei¯ora Ritter; AY875297, AY875231, AY875261, AY875322, land, Z. Portulacaria afra Jacq.;Ð, AY875219, AY875268,Ð, AY875368; Cult. YU; AY875354; Cult. B 166-54-83-20; Leuenberger et al. 3054, Brazil, B. Edwards 97, garden, YU. Pereskia bahiensis GuÈrke;Ð,Ð,Ð,Ð, AY875351; Cult. B 166-86-83-10; Quiabentia verticillata (Vaupel) A. Berger; AY875319, AY875239, Leuenberger et al. 3054, Brazil, B. AY875279, AY875340,Ð; Cult. B 154-12-74-80; Schwerdtfeger 22507, Pereskia bleo (Kunth) D.C.; AY875289, AY875227, AY875265, garden, B. AY875339,Ð; Cult. JBN; Edwards 13, garden, YU. Quiabentia zehntneri Britton & Rose; AY875307, AY875236, AY875274, Pereskia bleo (Kunth) D.C.;Ð,Ð,Ð,Ð, AY875359; Cult. B 277-01-88-80; AY875345, AY875372; Cult. B 163-09-88-30; Leuenberger et al. 3078, Schwerdtfeger 12678, garden, B. Brazil, B. Pereskia diaz-romeroana Cardenas;Ð,Ð,Ð,Ð, AY875353; Cult. B 039-02- Talinum paniculatum (Jacq.) Gaertn.; AY875316, AY875214, AY875252,Ð, 77-30; Rauh 40627, Bolivia, B. Ð; Cult. YU; Edwards 6, garden, YU. Pereskia grandifolia Howarth var. grandifolia; AY875298, AY875228, Talinum paniculatum (Jacq.) Gaertn.;Ð,Ð,Ð,Ð, AY015274; Cult. ZSS 93 AY875253, AY875325,Ð; Cult. NTG; Edwards 2, garden, YU. 1952; Martinez & Eggli 203, Mexico, ZSS. Pereskia grandifolia Howarth var. grandifolia;Ð,Ð,Ð,Ð, AY875362; Cult. Tephrocactus articulatus (Pfeiff.) Backbg.; AY875318, AY875248, B 038-04-77-80; Schwerdtfeger 12489, garden, B. AY875285, AY875347, AY875367; Cult. YU; Edwards 99, garden, YU.